How are fees charged on Kalshi and Polymarket?

Kalshi charges a taker fee of 7% × contracts × price × (1 − price), rounded up to the cent, per fill, on most series. Polymarket charges a taker fee of shares × category rate × price × (1 − price), with rates depending on the market category. Fees are charged on the trade, not on winnings.

Why do spreads and totals have more arbitrage opportunities than moneylines?

Spreads and totals move in discrete steps and are more sensitive to differences in models and information, leading to larger pricing gaps between venues. Moneylines are more stable and tightly clustered, so arbitrage is rarer.

Why Sports Spreads and Totals See More Arbitrage Opportunities Than Moneylines

By ArbMonster · Aug 19, 2026 · updated Sep 8, 2026

When you scan cross-market data—like ArbMonster does, comparing 20 sportsbooks with Kalshi and Polymarket—one pattern jumps out: spreads and totals in sports often generate more arbitrage opportunities than moneylines. But why? What makes these markets so much more prone to pricing mismatches?

In this article, we’ll break down the mechanics behind spreads, totals, and moneylines, show how they’re priced, and walk through worked examples with real-ish numbers. You’ll see exactly why spreads and totals are a hotspot for arbitrage, and how the math plays out in practice.

1. Market Structures: Moneylines vs. Spreads vs. Totals

Let’s start with the basics.

Moneyline: A bet on who wins. Example: Yankees -130 vs. Red Sox +120 (no spread).
Spread: A handicap applied to level the playing field. Example: Lakers -5.5 (-110), Celtics +5.5 (-110). Lakers must win by 6+; Celtics can win or lose by up to 5.
Total: A bet on the combined score. Example: Over 213.5 (-110), Under 213.5 (-110).

Prediction markets (like Kalshi and Polymarket) often list contracts mirroring these sportsbook bets: YES/NO on “Lakers win by 6+,” “Game totals over 213.5,” etc.

2. How Lines Are Set—and Why They Drift

Bookmakers use models, market input, and risk management to set lines. For moneylines, this is “who wins?”—a binary outcome. For spreads and totals, it’s a point estimate (margin of victory, total points), then a line is set so the odds are close to even (-110/-110 is standard).

But here’s the key:

Moneyline pricing is more stable. Most venues agree on who’s likely to win, and margins are small.
Spreads and totals move more—sometimes by a lot. Small differences in models, information, or betting action can shift the line by 0.5 or 1 point, creating large relative odds changes.

Example: NBA Spread Moves

Sportsbook A: Lakers -5.5 (-110)
Sportsbook B: Lakers -6.5 (-110)
Kalshi: YES on “Lakers win by 6+” at 45c (implied +122)

With a 1-point spread difference, the probability gap can be several percentage points—enough that after fees, arbitrage often appears.

3. The Math: How Small Line Differences Create Big Pricing Gaps

Let’s put numbers on this. Assume the true distribution of possible outcomes is roughly normal (as most models do for spreads/totals).

Step 1: Implied Probability from Spread

Suppose in an NBA game, the “Lakers -5.5” line at -110 implies:

Probability Lakers cover: 52.4% (from -110 odds)

If another book posts “Lakers -6.5” at -110, the probability they cover drops:

Probability Lakers cover -6.5: ~48.5%

So, a 1-point move reduces the win probability by about 4% (actual value depends on sport and scoring volatility).

Step 2: Arbitrage Example

Imagine:

Sportsbook A: Lakers -5.5 (-110) (YES side, implied 52.4%)
Kalshi: YES on “Lakers win by 6+” at 45c (implied 45%)

If you can back YES on Kalshi at 45c, and lay Lakers -5.5 at Sportsbook A for -110, there’s a 7.4% probability gap between markets.

Fees Matter

Kalshi charges a taker fee of 7% × contracts × price × (1 − price), rounded up to the cent, per fill, on most series. Sportsbook A’s odds have built-in vig.
After removing fees, if the gap remains, you may have an arbitrage opportunity.

Step 3: Why Moneylines Rarely Offer This

Moneylines for the same game tend to be tightly clustered:

Book A: Lakers -250 (implied 71.4%)
Book B: Lakers -245 (implied 71%)
Kalshi: YES at 71c

The difference is usually less than 1%. With fees, arbitrage rarely appears.

4. Why Spreads and Totals Are Prone to More Opportunities

4.1. More Line Variants

Spread and total lines move in discrete steps (0.5-point increments in NBA/NFL).
Different venues may be at “Lakers -5.5” vs. “Lakers -6.5” at the same time.

Each variant is effectively a different contract—so the pricing can diverge more.

4.2. Different Models, Different Moves

Bookmakers and prediction markets may use different models, data feeds, or react at different speeds to news. This affects where the line settles, especially for spreads/totals with less liquidity than moneylines.

4.3. Market Segmentation

Prediction markets sometimes list contracts at lines not available at sportsbooks anymore (old lines, stale data). Or, one sportsbook hangs a rogue number while others move. This is rarer for moneylines, where consensus is stronger.

5. Realistic Worked Example: NFL Total

Market Snapshot

Sportsbook A: Over 47.5 (-110), Under 47.5 (-110)
Sportsbook B: Over 48.5 (-110), Under 48.5 (-110)
Polymarket: YES on “Total points > 47.5” trading at 53c

Implied Probabilities

Over 47.5 at -110: 52.4%
Over 48.5 at -110: 48.5%
Polymarket YES at 53c: 53%

Can You Arbitrage?

Suppose you buy YES on “over 47.5” at 53c on Polymarket, and lay Under 47.5 at -110 at Sportsbook A. After Polymarket’s taker fee (≈4–7% × price × (1 − price), depending on category), your effective price is slightly higher than 53c (since the fee is charged on the trade, not on winnings).

If the sportsbook side’s implied probability is 52.4%, the pricing is nearly even post-fee. But if you find a sportsbook at +105 instead of -110, or Polymarket’s price is lower, the gap can widen enough to make a net-positive trade after fees.

This kind of edge almost never appears on moneylines, but shows up regularly on spreads and totals—especially in lower-liquidity or fast-moving markets.

6. How Automation Helps

With so many moving parts—dozens of books, both prediction markets, lines shifting by the minute—scanning for these opportunities manually is tough. That’s where a data service like ArbMonster comes in. It automates the hunt, netting out all fees, and only shows fillable opportunities (no phantom prices or shallow quotes).

You can learn more about the mechanics (including order book depth and fee math) at ArbMonster’s Learn page.

7. Key Takeaways

Spreads and totals move in steps; moneylines drift slowly.
A 1-point shift in a spread or total can create a 3–5% probability gap—enough to survive fees and create arbitrage.
Prediction markets and sportsbooks often disagree on spread/total lines, especially after news or in less-liquid games.
Automation is essential to catch fleeting, real opportunities—manual scanning likely misses them.

FAQ

Q1: Are spreads and totals always better for arbitrage than moneylines?
Not always, but the math shows gaps occur more often in spreads and totals due to line movement and variant pricing. Moneylines can have opportunities, but they’re rarer.

Q2: Do these gaps last long?
Usually not—once one venue moves, the window can close in minutes. Automation helps catch them before they vanish.

Q3: Can the same logic apply to other markets (like props or futures)?
Yes, any market with line variants or different pricing models can show arbitrage, but spreads and totals are especially prone due to discrete line moves and less consensus.

ArbMonster is a data service. Nothing here is financial or betting advice; markets carry risk, venues have age and jurisdiction restrictions, and you are responsible for verifying everything before acting.

← all articles · Glossary · How we measure